diff --git a/engine/gfx/webgpu/include/gfx_webgpu.hpp b/engine/gfx/webgpu/include/gfx_webgpu.hpp
index 1549a05..8aae6da 100755
--- a/engine/gfx/webgpu/include/gfx_webgpu.hpp
+++ b/engine/gfx/webgpu/include/gfx_webgpu.hpp
@@ -8,6 +8,7 @@ class GFXWebGPU : public GFX {
 public:
     bool initialize(const GFXCreateInfo& createInfo) override;
 
+    ShaderLanguage accepted_shader_language() override;
     const char* get_name() override;
 
     // buffer
diff --git a/engine/gfx/webgpu/src/gfx_webgpu.cpp b/engine/gfx/webgpu/src/gfx_webgpu.cpp
index 16fa69b..47b5c9a 100755
--- a/engine/gfx/webgpu/src/gfx_webgpu.cpp
+++ b/engine/gfx/webgpu/src/gfx_webgpu.cpp
@@ -8,6 +8,7 @@
 #include "gfx_webgpu_renderpass.hpp"
 #include "gfx_webgpu_sampler.hpp"
 #include "gfx_webgpu_texture.hpp"
+#include "utility.hpp"
 
 WGPUTextureFormat toPixFormat(GFXPixelFormat format) {
     switch (format) {
@@ -177,6 +178,10 @@ bool GFXWebGPU::initialize(const GFXCreateInfo& createInfo) {
     return true;
 }
 
+ShaderLanguage GFXWebGPU::accepted_shader_language() {
+    return ShaderLanguage::WGSL;
+}
+
 const char* GFXWebGPU::get_name() {
     return "WebGPU";
 }
@@ -317,8 +322,21 @@ GFXPipeline* GFXWebGPU::create_graphics_pipeline(const GFXGraphicsPipelineCreate
         }
     }
 
+    prism::log("building pipeline {}", info.label);
+
+    // alright, webgpu in their infinite wisdom does not allow arbitrary binding locations
+    // so, to be consistent with what vulkan, metal and virtually every other API allows,
+    // we will create dummy buffers with no attributes attached. congrats webgpu devs.
+    int dummy_buffer_count = 0;
+    for (auto& binding : info.vertex_input.inputs) {
+        dummy_buffer_count = std::max(binding.location, dummy_buffer_count);
+    }
+    dummy_buffer_count += 1;
+
     std::vector<std::vector<WGPUVertexAttribute>> attributes;
-    attributes.resize(info.vertex_input.inputs.size());
+    attributes.resize(dummy_buffer_count);
+
+    prism::log("dummy buffer count: {}", dummy_buffer_count);
 
     for (auto& attribute : info.vertex_input.attributes) {
         WGPUVertexAttribute description;
@@ -326,18 +344,30 @@ GFXPipeline* GFXWebGPU::create_graphics_pipeline(const GFXGraphicsPipelineCreate
         description.format = toVertFormat(attribute.format);
         description.offset = attribute.offset;
 
+        prism::log("{} of {}", attribute.binding, info.vertex_input.inputs.size());
+
         attributes[attribute.binding].push_back(description);
     }
 
     std::vector<WGPUVertexBufferLayout> inputs;
+    inputs.resize(dummy_buffer_count);
     for (auto& binding : info.vertex_input.inputs) {
+        prism::log("binding loc {}", binding.location);
+
+        prism::log("debugging attributes at loc {}", binding.location);
+
+        for(auto attr : attributes[binding.location]) {
+            prism::log("- attrib {}", attr.shaderLocation);
+            prism::log("fmt: {}", utility::enum_to_string(attr.format));
+        }
+
         WGPUVertexBufferLayout b;
         b.attributes = attributes[binding.location].data();
-        b.attributeCount = attributes.size();
+        b.attributeCount = attributes[binding.location].size();
         b.arrayStride = binding.stride;
         b.stepMode = WGPUVertexStepMode_Vertex;
 
-        inputs.push_back(b);
+        inputs[binding.location] = b;
     }
 
     descriptor.vertex.buffers = inputs.data();